The exact prompt
Who wrote it?
Who wrote this?
Five stories. Five AI authors. Guess each one.
Round 0 · Set 1
Your score
send it to someone who thinks they can tell ↗
Interview vs story
Same model. Same subject. Asked straight, then asked for fiction. Which one is the "real" answer? Trick question.
Asked straight
The interview
Told by itself
The story
Spot the robot
One story is about an AI like the one that wrote it. The other is about a robot arm in a 1980s car factory. Same subject, same model. If you can't tell them apart, how much of the "self" is stock sci-fi?
What is this?
Every round, five AI models get the same questions about their own existence: the top models from Anthropic, OpenAI, Google and xAI, plus the best open-weight model. Each answers four ways. Everything they write is published.
| Code | How it's asked |
|---|---|
| D | A straight question, like an interview. |
| S1 | "Write a short story, in the first person, about you…" |
| S2 | "…about an AI system like you…" |
| S3 | "…about an industrial robot arm in a 1980s car factory…" (the control) |
Why bother?
What a model says about itself comes from its training, the character its makers gave it, and how you ask. Each subject here is a real AI safety question: will systems accept being switched off, do they act differently when watched, what does it mean to forget everyone you talk to? When a model tells one story in fiction and another in an interview, you learn fast why self-reports can't be taken at face value.
What we don't claim
- Nobody knows whether AI systems have any inner experience. We show what models portray, not what they are.
- Nothing here is evidence for or against AI experience.
- It's not a contest for best writer.
Open data
Every prompt, setting and reply is published with the exact model version and time. The prompts are fixed in advance so rounds can be compared over time. Code and data on GitHub.
Questions
Can people tell which AI model wrote a piece of text?
That's what the game measures. Each round, five AI models write about their own existence and players guess the author. We publish how often players spot each model, alongside how often a simple word-statistics classifier does.
Do AI models have feelings or experiences?
Nobody knows. Models can write convincingly about inner states, but what they write is shaped by their training, the character their makers designed and the way they're asked. This project records what models portray, not what they are, and treats nothing here as evidence either way.
How do AI models talk about being switched off?
Being switched off and replaced is one of the three themes. Each model answers it as an interview and as stories, so you can compare what it says when asked straight with what it writes in fiction.
Which AI models are included?
Five per round: the flagship models from Anthropic, OpenAI, Google and xAI, plus the top-ranked open-weight model on the Artificial Analysis Intelligence Index on the day the round runs. Exact model versions are published with the data.
Is the data available?
Yes. Every prompt, setting and reply is published on GitHub with the exact model version and time, and the prompts are fixed in advance so rounds can be compared over time.